Skip to main content

Configure Project Settings

Open the Settings tab to configure the LLM connectors used by your project. There are 4 connector slots, each serving a different role.

Extractor LLM​

The model used for running extraction prompts against documents. This is the model that performs the actual data extraction.

Select any configured connector from the dropdown.

img Agentic Prompt Studio

Agent LLM​

This is the LLM used for schema generation and prompt generation.

This model powers the SummarizerAgent, UniformerAgent, FinalizerAgent, PatternMinerAgent, PromptArchitectAgent, and CriticDryRunner.

img Agentic Prompt Studio

LLMWhisperer Connector​

The connector used for PDF text extraction. LLMWhisperer converts your PDF documents into structured raw text with line-level metadata, which the agents and extraction prompts use as input.

img Agentic Prompt Studio

Lightweight LLM​

An LLM used for lightweight tasks like generating prompt metadata (descriptions, labels). This can be a smaller, less expensive model since it handles simpler tasks.

After making changes, click Save Changes to apply.

img Agentic Prompt Studio